The electroretinogram (ERG) has been studied for nearly a century. However, it has not yet developed into a powerful clinical tool. One explanation for this fact is that the elements of the retina interact in a complex nonlinear fashion, and most tests of retinal function rely on simple linear systems analysis techniques. The goal of this project is to characterize the ERG of human subjects for a variety of stimulus conditions using the nonlinear systems characterization procedure commonly called white-noise analysis. The first-and-second-order Wiener kernels computed for a white-noise stimulus reflect adaptation effects and other retinal mechanisms (Koblasz, 1976). During the first year of the project we will concentrate on improving the average signal-to-noise of the ERG measurement and will conduct a series of experiments aimed at resolving interactions between retinal color systems. Normal subjects and patients with known color anomalies will be used extensively. Portions of the temporal monlinear analysis will be conducted on-line using a TTL processor recently constructed. During the second and third years of the project, we will consider spatial interactions. We will first attempt to characterize the ERG for small focal stimuli using counterphase techniques without retinal stabilization. We will then provide two spatially separated stimuli and hopefully resolve interactions between the two stimulated regions of the retina. Most of these experiments will be conducted on normal subjects, but several patients with retinal detachment will be tested before and after reattachment. If these experiments are successful, then we will shrink the focal stimuli to smaller dimensions, requiring retinal stabilization. The final set of experiments will focus on spatial interactions of small dimensions and will be conducted only on normal subjects.